Within the past 10 years, there have been major improvements in management of patients having the more advanced stages of diabetic retinopathy. However, relatively little is known about the course of disease from its earliest manifestations to the advanced stages. Both the Sorbinol Retinopathy Trial (a completed clinical trial) and the Diabetes Complications and Control Trial (DCCT) (an ongoing clinical trial) have used changes in the grade of early diabetic retinopathy as an outcome measure without substantial evidence on the clinical significance of the timing of these changes. Long-term longitudinal data on a diabetic population are required to address this issue. The purpose of this application is to use an existing data set to analyze risk factors for and the natural course of diabetic retinopathy in Type I (insulin-dependent) diabetics. This dataset has an annual followup (median 8 years) information on clinical and retinal status on a cohort of 269 Type I diabetics. The analyses proposed will answer: 1) does the duration of disease at which advanced retinopathy appears depend on the duration at which early background retinopathy appears? 2) does the occurrence of microalbuminuria, a very early indicator of renal damage, act as a predictor for retinal disease? 3) does less stringent control of glycemic levels result in more fluctuation (appearance and disappearance) in background retinopathy? and 4) does changing glycemic control over time affect the risk of progression of retinopathy? This dataset is uniquely able to explore such questions, which have important clinical implications for other studies of diabetic retinopathy.